That’s already been done for assisting vision impaired voters.
Correct, we’ve raised this issue numerous times in the past and the AEC and others always wheedle their way out of it by claiming “commercial in confidence” BS.
Not quite all. There was one piece of software used in the ACT about 12 or 13 years ago that the company who got the tender released the full source code voluntarily and before anyone asked. When asked why by American journalists (it was around the same time as the Diebold stuff), the lead developer said that it was more important for a transparent democracy and doing otherwise would be unethical. IIRC, they were not a regularly FOSS development house, they just viewed voting software as different and with special requirements (that wereb’t part of the tender either).
Can’t remember the name of the company, though. They weren’t very big.
Really? A little earlier things looked very much like playing blockchain buzzword bingo.
Not only that, but those dtrengths it does have are in the distributed processing and participation. Running a federal, state or local election, however, has a considerable number of legal requirements and even if it were to be adopted. the situation would still necessitate either a centralisation of the computing resources with the AEC or a vetting process for third party participation (including hardware vetting).
If, however, you were leaning more towards not only distributing the load, but also using it to facilitate casting votes through direct participation in maintaining that blockchain (i.e. running a full node), then you start to skate into the very tricky area of needing to maintain the secrecy of the ballot, while still confirming each person only casts one vote.
It’s not designed for this, but engineers everywhere have found their newest, favourite hammer and now everything looks like a nail.
The advantage with the current system of scanning physical ballots, though, is that if there are inconsistencies with coding the ballots in then the original can be double-checked. The same process was used with the 2011 Census and it’s actually pretty decent. The only inverse way to do it is to complete ballots on a voting machine, print the ballot and submit that; with the printed ballot being the official vote to be confirmed and the electronic record just being available as a convenience.
Yes. Plus it should all be publicly accessible and open source.